CN103957246B - Dynamic load balancing method based on tenant's perception and system - Google Patents
Dynamic load balancing method based on tenant's perception and system Download PDFInfo
- Publication number
- CN103957246B CN103957246B CN201410163481.8A CN201410163481A CN103957246B CN 103957246 B CN103957246 B CN 103957246B CN 201410163481 A CN201410163481 A CN 201410163481A CN 103957246 B CN103957246 B CN 103957246B
- Authority
- CN
- China
- Prior art keywords
- request
- application server
- tenant
- server
- cluster
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of dynamic load balancing method based on tenant's perception and system, the method comprising the steps of：Receive the request of client, and extract the tenant's id information in this request；According to described tenant's id information, described request is sent in corresponding server cluster, and calculates the number of request processing in this server cluster；Judge whether described number of request is more than the first predetermined threshold value；If so, then pass through to dock the new application server of IaaS platform establishment；The new application server creating is added in described server cluster, configuration file required for being loaded from configuration file warehouse by automatic script is to the new application server creating, and trigger simultaneously operating, code in central code storage is checked out the application server of new establishment, and after the deployment of application server completing newly to create, forward the request on the application server that newly creates.The method of the present invention and system improve the utilization ratio of server resource.
Description
Technical field
The present invention relates to communication technical field, more particularly to a kind of dynamic load balancing method based on tenant's perception with
And a kind of dynamic load balancing system based on tenant's perception.
Background technology
Traditional load-balancing method, is to will determine that by unified configuration file the request of module is routed to specify
In one server cluster, the server in cluster needs to reserve, and completes to install and portion during the operation of server application needs
Administration's application.
The shortcoming of such scheme is it is clear that the quantity of server needs enterprise to client after completing exploitation and testing
Make assessment with the request amount of system, to tackle the pressure of system.But in a large amount of medium-sized and small enterprises in practice, the pressure of system
Only peak just occurs in extremely individual other period, enterprise, in order to tackle a small amount of peak, generally requires to be equipped with for application service
The server of redundancy.This way causes the significant wastage of server resource, for medium-sized and small enterprises, server
Resource more should stress efficient utilization.
Content of the invention
Based on this, the present invention provides a kind of dynamic load balancing method based on tenant's perception and system, it is possible to increase clothes
The utilization ratio of business device resource.
For achieving the above object, the present invention adopts the following technical scheme that：
A kind of dynamic load balancing method based on tenant's perception, comprises the following steps：
Receive the request of client, and extract the tenant's id information in this request；
According to described tenant's id information, described request is sent in corresponding server cluster, and calculates this server set
The number of request processing in group；
Judge whether described number of request is more than the first predetermined threshold value；
If so, then pass through to dock the new application server of IaaS platform establishment；
The new application server creating is added in described server cluster, by automatic script from configuration file warehouse
Configuration file required for middle loading is to the new application server creating, and triggers simultaneously operating, by central code storage
Code checks out the application server of new establishment, and after the deployment of application server completing newly to create, will described ask
Ask on the application server being forwarded to new establishment.
A kind of dynamic load balancing system based on tenant's perception, including：
Extraction module, for receiving the request of client, and extracts the tenant's id information in this request；
Computing module, for described request being sent in corresponding server cluster according to described tenant's id information, and
Calculate the number of request processing in this server cluster；
First judge module, for judging whether described number of request is more than the first predetermined threshold value；
Creation module, for the judged result in described first judge module for, in the case of being, being put down by docking IaaS
Platform creates new application server；
Expansion module, for being added to the new application server creating in described server cluster, by automatic script
Configuration file required for loading from configuration file warehouse is to the new application server creating, and triggers simultaneously operating, in inciting somebody to action
Code in the code storage of centre checks out the application server of new establishment, and completes the portion of application server that newly creates
After administration, forward the request on the application server that newly creates.
By above scheme as can be seen that a kind of dynamic load balancing method based on tenant's perception of the present invention and system,
Automatically different requests can be forwarded in different server clusters, and specific clothes are forwarded to according to tenant's id information
In business device cluster, this load balancing based on tenant's perception, based on the assemblage classification of tenant ID, can preferably manage difference
The tenant of request amount and application；And present invention achieves the docking of application server system and IaaS platform, need not reserve should
With the quantity of server, after number of request reaches early warning value, by being created that new application server simultaneously in IaaS platform
Automatically application is deployed in the resilient expansion that can achieve application server thereon, drastically increases the utilization of server resource
Efficiency.
Brief description
Fig. 1 is a kind of schematic flow sheet of the dynamic load balancing method based on tenant's perception of the present invention；
Fig. 2 is the load balancing schematic diagram based on tenant's perception in the embodiment of the present invention；
Fig. 3 is the schematic diagram realizing dynamic load leveling in the embodiment of the present invention；
Fig. 4 is the dynamic calculation number of request mentality of designing figure in the embodiment of the present invention；
Fig. 5 is the structural representation of the group system in the embodiment of the present invention；
Fig. 6 is a kind of structural representation of the dynamic load balancing system based on tenant's perception of the present invention.
Specific embodiment
In order that the objects, technical solutions and advantages of the present invention become more apparent, below in conjunction with drawings and Examples, right
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only in order to explain the present invention, and
It is not used in the restriction present invention.
Embodiment one
In the production environment of enterprise, Web application system development needs the web application server of stability and high efficiency.In cloud meter
In the epoch calculated, advocate resource-sharing, distribution according to need.The proposition of the particularly application engine such as google app engine, application support
Pipe has begun to popular.The realization of application engine in the present invention, by building a resilient expansion and the application server shrinking
Cluster externally provides unified AH service.The application system of enterprise development is disposed by this application hosted platform,
Runtime environment required for application system is run is provided by application engine.The dynamic representation of system is in a publicly-owned cloud environment
In, all applications may be all the container of a shared application service, and built-in dynamic load-balancing algorithm reaches pre- in number of request
After alert value, copy new application server, server is automatically added to original cluster, shunting request.Platform is with registered enterprise
(or user), as tenant, shares cluster container between tenant, accomplish application isolation simultaneously；Platform provides unified back-stage management to enter
Mouthful, distributed application server cluster, central management warehouse database, and the load balancing of tenant's perception, based on tenant
The assemblage classification of ID, can more preferably manage tenant and the application of different request amount.
Application server cluster adopts the framework of principal and subordinate's loose coupling, and management node is responsible for the management of cluster
(management), execution node is responsible for the response (execution) of user's request, negative by the customization of configurable xml document
Carry the code of practice of equalizing system, system judges the load shape of the application server in cluster using Dynamic Load-balancing Algorithm
Condition, when load reaches early warning value, by docking IaaS (Infrastructure as a Service, infrastructure service)
Platform creates new application server instance application deployment, thus will determine that the request of module is forwarded to the application clothes of new establishment
On business device, then it is shown to judge module returning result.After number of request declines, again can be by nullifying in IaaS platform
Certain applications server Resource recovery.This characteristic can greatly improve stability and the high availability of application system, and
The utilization rate of server resource.
Shown in Figure 1, a kind of dynamic load balancing method based on tenant's perception, comprise the following steps：
Step S101, receives the request of client, and extracts the tenant's id information in this request.General, tenant ID believes
Breath can be transmitted by the form of http header file.
Step S102, according to described tenant's id information, described request is sent in corresponding server cluster, and calculates
The number of request processing in this server cluster.
As shown in Fig. 2 the initialization of server cluster specifies different tenant's application distribution in different sub-cluster,
Apply and carry tenant's id information in central code storage, when being deployed in different application sub-cluster, by tenant ID
Information transmission is to SiteServer LBS.When user accesses application, load balancing, according to the target tenant ID extracting, will be asked
It is forwarded in specific server cluster.The method is conducive in the environment of multi-tenant, and the application of Some Enterprises trustship may
Need to expend great amount of hardware resources, number of request is huge, and this kind of application can be separately configured a cluster, to process user's request,
The application isolation all less with other resource requirements and request.
Step S103, judges whether described number of request is more than the first predetermined threshold value；If then entering step S104, otherwise not
Need to process.
Step S104, if the judged result of step S103 is yes, creates new application service by docking IaaS platform
Device.
Step S105, the new application server creating is added in described server cluster, by automatic script from joining
Put the configuration file required for loading in file repository to the application server of new establishment, and trigger simultaneously operating, by central generation
Code in code warehouse checks out the application server of new establishment, and completes the deployment of application server that newly creates
Afterwards, forward the request on the application server that newly creates.
Present invention achieves dynamic load leveling.As shown in figure 3, so-called dynamic, refer to achieve application server system
System and the docking of IaaS, need not reserve the quantity of application server, after number of request reaches early warning value, by calling JCloud
Interface is created that virtual machine in IaaS environment, after the masterplate of virtual machine creates, by the script of automatic deployment, installs application
Application is deployed in thereon, by load equalizer by the operation of server after start-up execution code synchronisation automatically
Too high request is forwarded in the application server instance of new establishment.And, the present invention is being created that new application server
Afterwards, shrink process can also be carried out according to real-time loading condition.As a preferable embodiment, the present invention is calculating clothes
After the number of request processing in business device cluster, step can also be included as follows：Judge whether described number of request is less than second
Predetermined threshold value；If so, then find out in IaaS platform and load minimum application server and be unregistered.
In addition, as a preferable embodiment, the process of the number of request processing in described calculation server cluster
Specifically can include as follows：
When reception receives a request, it is this request one unique ID of distribution by inputting moderator, and by this ID
Put in messaging list；
If described request is detected meeting with a response from application service, by exporting the ID that this is asked by moderator from described
Deleted in messaging list；
By calculate the ID quantity in current message list obtain described in the number of request that processing.
It is illustrated in figure 4 the mentality of designing of dynamic calculation number of request in the present invention.System includes inputting message queue, input
Message queue and three assemblies of analyzer.When system receives a request, input moderator can distribute one uniquely for message
ID, message id is put in messaging list, if request from service meet with a response, output moderator can be by the ID of this message
Delete from messaging list, by such mode, the actual number having grasped the request that background service is being processed of messaging list
Amount.And, analyzer passes through the length that timing calculates messaging list, and this length is compared with the threshold value in configuration,
Once reaching extension or the scope shunk, analyzer will call the interface of JCloud, and reality is being processed with the application clothes of request
Business device expands or contracts operation.
As a preferable embodiment, a kind of dynamic load balancing method based on tenant's perception of the present invention, also may be used
To comprise the steps：It is managed node and execution node separating treatment in described server cluster；Wherein, described management
Node is used for uploading of engineering code and downloads, applies deployment and the life cycle management of application；Described execution node is used for executing
And respond the request of judge module.As shown in figure 5, the present invention makes full use of application management server node and execution node is realized
Front and back end separates and keeps light weight and high efficiency, and SiteServer LBS realizes stability and the High Availabitity of application server system
Property, the drawbacks of efficiently solving traditional server single node and work.
In addition, as a preferable embodiment, the new application server creating is added to described service in the present invention
Process in device cluster specifically may include steps of：By the setting of configuration file in the application server of described new establishment
Put, go out to add the well-known node that server cluster used and its IP address according to wka protocol identification, and according to described IP address
Find well-known node, the application server of described new establishment is added in the server cluster corresponding to this well-known node.
As shown in figure 5, specifying the essential information of the cluster of foundation in load equalizer, application server node is passed through to join
Put the setting of file, go out to add the well-known node (well-known-node) that cluster used and needs according to wka protocol identification
The cluster adding.The deployment that tenant is applied from management node, and trigger the application sync agent on execution node, execute section
Point detects code automatically from code storage, completes to apply the deployment on execution node；When request is through load balancing
Wait, forward requests to execute on node by the mapping of port, node completes to the corresponding of request and returns to judge module.
Embodiment two
The present invention relates to web application server cluster technology, the particularly architecture design of application server cluster, and
The realization of the configurable dynamic load leveling device based on tenant's perception.Present invention is primarily concerned with 2 points：One is the dynamic of system
Load balance ability, two is in the cluster of load balancing, realizes the ability of tenant's perception.Dynamic load leveling device and load are all
Weighing apparatus is equally located at whole system foremost, receives all requests from judge module, the request receiving is routed to
In the execution node of the sub-cluster specified, and by executing node, request is responded；Difference is dynamic load leveling
Device needs to extract warning data and algorithm, the control example in cluster being extended and shrinking by reading configuration file.
According to these technical requirements, concrete methods of realizing and step are as follows：
First, the dynamic load leveling device of the present invention needs to realize two core interface.One is the technical interface of cluster.Two are
The interface of dynamic expansion application server.
1) ultimate principle realizing cluster interface is as follows：Create well-known node first, well-known node belongs to certain server
Cluster, other nodes pass through to state this well-known node and its IP address in configuration, just can be found according to IP after startup
Well-known node, and be added in its corresponding cluster.In app engine, load equalizer is this well-known node,
Application server just loads configuration after starting, and is added in cluster.
2) need the number of request of finger daemon calculation server example in load equalizer.Realize principle as shown in Figure 4.
After reaching early warning value or being less than lower limit, dynamic socket needs extension or nullifies application server instance.Dynamic expansion and
The interface shrinking is realized mainly using JCloud (a set of general purpose I aaS management interface of increasing income), compatible Openstack, Amazon
The IaaS environment such as EC2.
3) in static state setting cluster, the maximum quantity of application server is 5, and minima is 1 guarantee system operation (max
Instance=5；Min instance=1；Max instance=-1 represents that maximum is unrestricted).
4) the load balancing configuration realized based on tenant's perception, tenant registers in application engine will be using unique
TanantID is identified, and carries out the configuration of user's perception in loadbalancer.xml using this ID.Two clusters of setting
And dispose identical application, process the tenant request of different tenant's ID scopes.Domain name
Domain.gcidesign.appengine1.com and domain.gcidesign.appengine2.com, in the configuration of cluster
In file, understand the different tenant ID handled by two clusters of explicit definition, such as cluster
1domain.gcidesign.appengine1.com can only process the request of the tenant in 1-10 for the tenant ID；Cluster
2domain.gcidesign.appengine2.com processes the request of other tenants remaining；The ID of tenant can be by http head literary composition
The form of part passes to load equalizer, and then processes different requests respectively.
5) how the node in cluster defined in dynamic load leveling device extends and shrinks.As initial dynamic algorithm, mesh
Front algorithm is relatively simple, can improve in conjunction with anticipation scheduling algorithm, and the number of request that in the present invention, real-time calculating is being processed is simultaneously
Carry out the realization of dynamic expansion according to number of request.Configuration file definition triggering calculating cycle T, unit interval receive the maximum of request
The parameters such as numerical value RPS, high thresholdhigh early warning rate AUR, lower limit prompting rate ALR.Dynamic Load-balancing Algorithm is that one kind expands in advance
Exhibition, postpones the cluster thinking shunk.When computation requests number reaches individual node largest request number 80%, newly-increased request
Start to be forwarded on other node, in request, less than individual node largest request number 20%* weight factor, (weight factor is current
Only represent the sensitivity to shrink request number) when, this node will be canceled, and request will be recovered in other nodes.
Upper limit threshold T hreshold1=RPS*T*AUR；
Lower limit threshold T hreshold2=RPS*T*ALR*FACTOR；
Wherein FACTOR parameter is used for determining the sensitivity of lower limit, between value 0-1.Computation requests number is connect using queue
Receive the Twice requests number within the nearest stipulated time of load equalizer, then calculate its meansigma methods.Example such as following table：
| RQ cycle | 1 | 2 | 3 | 6 | 7 | 8 | 9 |
| Online request | 10 | 1 | 250 | 200 | 300 | 15 | 0 |
Assume the result upper limit threshold value 200 of configuration；Lower limit is 30.
After system start-up, in first time interval, receive 10 requests, queue length is 1, meansigma methodss are 5<200；System
System example keeps constant；
The next time receives 1 request, queue length=2, meansigma methodss=5.5<200, example keeps constant；
The next time, queue length=2, meansigma methodss=125.5<200, example keeps constant；
Next time average=225>200, start extension example, create new application server instance；
Next time average=250<2*200, system example number keeps constant；
Next time average 30*1<157.5<2*200, system example number keeps constant；
Next time average 7.5<30*1, the operation that system execution is shunk, the minimum machine of de-registration request.
The form that node thresholding forwards, when load equalizer receives request, acquiescence is forwarded to one of
On execution node, in load equalizer, there is a finger daemon, timing calculates the number of request of every machine in current whole cluster.Work as team
When the result of column count is more than load balancing forwarding threshold value, load equalizer will embody dynamic ability.Guard into
Journey can be created that application server in IaaS environment and add in original cluster by triggering example controller calling interface,
Application server is loaded from original application server by automatic script after starting and is configured to locally, and triggers synchronous behaviour
Make, the code in code storage is checked out local, now application server has completed disposing and running of application, hereafter bears
Carry equalizer just to start to make requests on forwarding to newly-built execution node in cluster.
6) configuration load equilibrium externally exposes the domain name addresses of service, so that after the new application server instance creating starts
Can find and add cluster, the execution node of all new establishments is all added to the cluster of acquiescence by the domain name of load equalizer
In.
2nd, management node separates with execution node, and concrete operation step is as follows：
1) there are two subdomains in dynamic load leveling device cluster, management domain, as control node, controls triggering code same
Step signal, the execution node in execution domain is that dynamic load leveling device needs extension and the subdomain shrinking, and subdomain passes through mgt respectively
It is labeled with worker；
2) corresponding option, the configuration letter of each node that system is used are configured in each application server instance
Breath, is all left on configuration file main frame by the good xml or properties file of configured in advance.The new execution node creating
Start all automatic running automatized scripts after finishing to configure with properties：
A. application server uses apache tribe Clustering, using wka agreement；
B. the new node creating is specified to be execution node, subdomain entitled " worker "：
C. by configuring well-known node for load equalizer, it is added in cluster；
D. the link address in configuration code warehouse, username and password, prepares to check out code；
E. after being added to cluster, receive the request of management node in management domain, triggering checks out the operation of code；
F. start application server.
3rd, dispose central code administration warehouse, to realize executing the application code synchronization between node.Central code storage
Configured using SVN it is intended that project is in the address of SVN, username and password.The ID of implicit tenant in SVN warehouse, can be true
Protect the ability of tenant's perception.After new execution node creates, SVN is installed by automatic script, and from central code storage
Check out application code.
It should be noted that there being two warehouses in the present invention：Configuration file warehouse and central code storage.By JCloud
After the virtual machine instance that interface creates in the IaaS environment such as Openstack starts, by the agent process in new node
(agent process) and automatic running script load required configuration file to add cluster in configuration file warehouse,
And realize executing the application synchronization between node by checking out the code of central code storage.
The present invention is the application server cluster of a multi-tenant, belongs to cloud computing platform and services (Platform-as-
A-Service part of functions) is realized.Weblication and web service, platform can be disposed by application engine
Environment when offer program is run.Enterprise customer's registration on platform in the form of tenant uses, all Application share container sections
The expense of resource-saving, process isolation ensures the safety of respective application system.
Platform using the present invention, it is possible to achieve the dynamic load leveling based on service and tenant's perception, on the one hand realizes
The ability of load balancing, automatically different requests is forwarded in different application clusters, and the id information according to tenant, turns
It is dealt in specific tenant's cluster；Based on the load balancing of tenant's perception, it is more beneficial for the hardware money of reasonable distribution application engine
Source.On the other hand, by way of configuring, resilient expansion and contraction can be carried out to application system according to real-time loading condition,
It is integrated to a PaaS and IaaS example well.
Corresponding with a kind of above-mentioned dynamic load balancing method based on tenant's perception, the present invention also provides a kind of being based on to rent
The dynamic load balancing system of family perception, as shown in fig. 6, include：
Extraction module 101, for receiving the request of client, and extracts the tenant's id information in this request；
Computing module 102, for described request being sent in corresponding server cluster according to described tenant's id information,
And calculate the number of request processing in this server cluster；
First judge module 103, for judging whether described number of request is more than the first predetermined threshold value；
Creation module 104, for described first judge module judged result be in the case of, by dock IaaS
Platform creates new application server；
Expansion module 105, for being added to the new application server creating in described server cluster, by automatic foot
This configuration file required for loading from configuration file warehouse is to the new application server creating, and triggers simultaneously operating, will
Code in central code storage checks out the application server of new establishment, and in the application server completing newly to create
After deployment, forward the request on the application server that newly creates.
As a preferable embodiment, a kind of dynamic load balancing system based on tenant's perception of the present invention is acceptable
Including：
Second judge module, for judging whether described number of request is less than the second predetermined threshold value；
Shrink module, for the judged result in described second judge module for, in the case of being, finding out in IaaS platform
Load minimum application server and be unregistered.
It is appreciated that term " first " used in the present invention, " second " etc. can herein be used for describing various elements,
But these elements should not be limited by these terms.These terms are only used for distinguishing first element with another element.Citing comes
Say, without departing from the scope of the invention, the first judge module can be referred to as the second judge module, and similarly,
Second judge module can be referred to as the first judge module.First judge module and the second judge module both judge modules,
But it is not same judge module.
As a preferable embodiment, described computing module can include：
ID distribute module, for when reception receives a request, distributing one only by inputting moderator for this request
One ID, and this ID is put in messaging list；
ID removing module, if meet with a response from application service for described request is detected, by exporting moderator
This ID asking is deleted from described messaging list；
ID number calculating section, for by calculate current message list in ID quantity obtain described in processing please
Ask several.
As a preferable embodiment, a kind of dynamic load balancing system based on tenant's perception of the present invention is acceptable
Including：
Node separation module, for being managed node and execution node separating treatment in described server cluster；Its
In, described management node is used for uploading of engineering code and downloads, applies deployment and the life cycle management of application；Described execute section
Point is for executing and responding the request of judge module.
As a preferable embodiment, described expansion module can include：
Cluster adds module, for the setting of configuration file in the application server by described new establishment, is assisted according to wka
Discuss the well-known node and its IP address identifying that addition server cluster is used, and well-known section is found according to described IP address
Point, the application server of described new establishment is added in the server cluster corresponding to this well-known node.
A kind of other technical characteristics of above-mentioned dynamic load balancing system based on tenant's perception and a kind of base of the present invention
Identical in the dynamic load balancing method of tenant's perception, it will not go into details herein.
By above scheme as can be seen that a kind of dynamic load balancing method based on tenant's perception of the present invention and system,
Automatically different requests can be forwarded in different server clusters, and specific clothes are forwarded to according to tenant's id information
In business device cluster, this load balancing based on tenant's perception, based on the assemblage classification of tenant ID, can preferably manage difference
The tenant of request amount and application；And present invention achieves the docking of application server system and IaaS platform, need not reserve should
With the quantity of server, after number of request reaches early warning value, by being created that new application server simultaneously in IaaS platform
Automatically application is deployed in the resilient expansion that can achieve application server thereon, drastically increases the utilization of server resource
Efficiency.
The description of specific distinct unless the context otherwise, the element in the present invention and assembly, the shape that quantity both can be single
Formula exists it is also possible to presented in multiple, the present invention is not defined to this.Although the step in the present invention is entered with label
Go arrangement, but be not used to limit the precedence of step, unless expressly stated the holding of the order of step or certain step
Based on row needs other steps, otherwise the relative rank of step is adjustable in.
Embodiment described above only have expressed the several embodiments of the present invention, and its description is more concrete and detailed, but simultaneously
Therefore the restriction to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, some deformation can also be made and improve, these broadly fall into the guarantor of the present invention
Shield scope.Therefore, the protection domain of patent of the present invention should be defined by claims.
Claims (8)
1. a kind of dynamic load balancing method based on tenant's perception is it is characterised in that comprise the following steps：
Receive the request of client, and extract the tenant's id information in this request；
According to described tenant's id information, described request is sent in corresponding server cluster, and calculates in this server cluster
The number of request processing；
Judge whether described number of request is more than the first predetermined threshold value；
If so, then pass through to dock the new application server of IaaS platform establishment；
The new application server creating is added in described server cluster, is added from configuration file warehouse by automatic script
Configuration file required for carrying is to the new application server creating, and triggers simultaneously operating, by the code in central code storage
Check out the application server of new establishment, and after the deployment of application server completing newly to create, described request is turned
It is dealt on the application server of new establishment；
It is managed node and execution node separating treatment in described server cluster；Wherein, described management node is used for work
Uploading of range code is downloaded, application is disposed and the life cycle management of application；Described execution node is used for executing and responds judgement
The request of module.
2. the dynamic load balancing method based on tenant's perception according to claim 1 is it is characterised in that calculating service
After the number of request processing in device cluster, also include step：
Judge whether described number of request is less than the second predetermined threshold value；
If so, then find out in IaaS platform and load minimum application server and be unregistered.
3. the dynamic load balancing method based on tenant's perception according to claim 2 is it is characterised in that described calculating takes
The process of the number of request processing in business device cluster includes：
When reception receives a request, distribute a unique ID by inputting moderator for this request, and this ID is put into
In messaging list；
If described request is detected meeting with a response from application service, by exporting the ID that this is asked by moderator from described message
Deleted in list；
By calculate the ID quantity in current message list obtain described in the number of request that processing.
4. according to claim 1-2 any one based on tenant perception dynamic load balancing method it is characterised in that
The process that the new application server creating is added in described server cluster includes：
By the setting of configuration file in the application server of described new establishment, go out to add server set according to wka protocol identification
Well-known node and its IP address that group is used, and well-known node is found according to described IP address, by the application of described new establishment
Server is added in this server cluster corresponding to well-known node.
5. a kind of dynamic load balancing system based on tenant's perception is it is characterised in that include：
Extraction module, for receiving the request of client, and extracts the tenant's id information in this request；
Computing module, for described request being sent in corresponding server cluster according to described tenant's id information, and calculates
The number of request processing in this server cluster；
First judge module, for judging whether described number of request is more than the first predetermined threshold value；
Creation module, for the judged result in described first judge module for, in the case of being, being created by docking IaaS platform
Build new application server；
Expansion module, for being added to the new application server creating in described server cluster, by automatic script from joining
Put the configuration file required for loading in file repository to the application server of new establishment, and trigger simultaneously operating, by central generation
Code in code warehouse checks out the application server of new establishment, and completes the deployment of application server that newly creates
Afterwards, forward the request on the application server that newly creates；
Node separation module, for being managed node and execution node separating treatment in described server cluster；Wherein, institute
State management node to download for uploading of engineering code, apply deployment and the life cycle management of application；Described execution node is used
In the request executing and responding judge module.
6. the dynamic load balancing system based on tenant's perception according to claim 5 is it is characterised in that also include：
Second judge module, for judging whether described number of request is less than the second predetermined threshold value；
Shrink module, for the judged result in described second judge module in the case of being, finding out load in IaaS platform
Minimum application server is simultaneously unregistered.
7. the dynamic load balancing system based on tenant's perception according to claim 6 is it is characterised in that described calculating mould
Block includes：
ID distribute module, for when reception receives a request, distributing one uniquely by inputting moderator for this request
ID, and this ID is put in messaging list；
ID removing module, if meeting with a response from application service for described request is detected, should by exporting moderator
The ID of request is deleted from described messaging list；
ID number calculating section, for by calculate the ID quantity in current message list obtain described in the request that processing
Number.
8. according to claim 5-7 any one based on tenant perception dynamic load balancing system it is characterised in that
Described expansion module includes：
Cluster adds module, for the setting of configuration file in the application server by described new establishment, is known according to wka agreement
Do not go out to add the well-known node that server cluster used and its IP address, and well-known node is found according to described IP address, will
The application server of described new establishment is added in this server cluster corresponding to well-known node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410163481.8A CN103957246B (en) | 2014-04-22 | 2014-04-22 | Dynamic load balancing method based on tenant's perception and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410163481.8A CN103957246B (en) | 2014-04-22 | 2014-04-22 | Dynamic load balancing method based on tenant's perception and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103957246A CN103957246A (en) | 2014-07-30 |
| CN103957246B true CN103957246B (en) | 2017-03-01 |
Family
ID=51334469
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410163481.8A Active CN103957246B (en) | 2014-04-22 | 2014-04-22 | Dynamic load balancing method based on tenant's perception and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103957246B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104158911A (en) * | 2014-09-01 | 2014-11-19 | 青岛海信电器股份有限公司 | Server cluster-oriented control method and system |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106209666B (en) * | 2015-05-07 | 2020-06-02 | 中兴通讯股份有限公司 | Link multiplexing method and system based on load balancer |
| CN106302570A (en) * | 2015-05-14 | 2017-01-04 | 阿里巴巴集团控股有限公司 | A kind of request processing method and device |
| CN106571943A (en) * | 2015-10-10 | 2017-04-19 | 北京国双科技有限公司 | Distributed-type configuration cluster capacity-expanding method and device |
| CN105868011A (en) * | 2016-03-29 | 2016-08-17 | 上海斐讯数据通信技术有限公司 | Method and device for improving running efficiency of software system |
| US10153941B2 (en) * | 2016-05-17 | 2018-12-11 | Microsoft Technology Licensing, Llc | Distributed operational control in computing systems |
| CN105959410B (en) * | 2016-06-29 | 2019-06-04 | 深圳中兴网信科技有限公司 | The load-balancing method and SiteServer LBS of server cluster |
| CN106203810B (en) * | 2016-07-01 | 2020-05-19 | 上海中传网络技术股份有限公司 | CRM data processing method based on cloud platform |
| CN108093026B (en) * | 2017-11-17 | 2020-04-07 | 阿里巴巴集团控股有限公司 | Method and device for processing multi-tenant request |
| CN107995286B (en) * | 2017-11-30 | 2021-03-02 | 平安科技（深圳）有限公司 | Automatic service starting and stopping method based on dubbo platform, server and storage medium |
| CN110865881A (en) * | 2018-08-27 | 2020-03-06 | 中移(苏州)软件技术有限公司 | Resource scheduling method and device |
| CN109582451A (en) * | 2018-11-21 | 2019-04-05 | 金色熊猫有限公司 | Method for scheduling task, system, equipment and readable medium |
| CN111240749B (en) * | 2018-11-28 | 2023-07-21 | 中国移动通信集团浙江有限公司 | Suspending control method, device, equipment and storage medium of instance in cluster system |
| CN109614241B (en) * | 2018-12-17 | 2024-03-12 | 杭州玳数科技有限公司 | Method and system for realizing multi-cluster multi-tenant resource isolation based on Yarn queue |
| CN109412878A (en) * | 2019-01-16 | 2019-03-01 | 紫光云数科技有限公司 | Multi-tenant service access implementation method, device and electronic equipment |
| CN110138862B (en) * | 2019-05-16 | 2022-03-04 | 北京字节跳动网络技术有限公司 | Service processing method, device, gateway equipment and medium based on application service |
| CN110633151B (en) * | 2019-09-20 | 2022-09-02 | 北京小米移动软件有限公司 | Method, device and storage medium for balancing distributed message issuing cluster partitions |
| CN111092816B (en) * | 2019-11-26 | 2022-02-15 | 金蝶软件（中国）有限公司 | Gateway dynamic routing method, gateway dynamic routing device, computer equipment and storage medium |
| CN113296931A (en) * | 2020-07-14 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Resource control method, system, computing device and storage medium |
| CN113765966A (en) * | 2020-09-01 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Load balancing method and device |
| CN112468313B (en) * | 2020-10-28 | 2023-07-07 | 北京思特奇信息技术股份有限公司 | WebLogic cluster rapid deployment method and system |
| CN114095419B (en) * | 2021-11-12 | 2023-11-28 | 软通动力信息技术(集团)股份有限公司 | Cluster routing method, device, system and storage medium |
| CN114679412B (en) * | 2022-04-19 | 2024-05-14 | 浪潮卓数大数据产业发展有限公司 | Method, device, equipment and medium for forwarding traffic to service node |
| CN115604244A (en) * | 2022-07-01 | 2023-01-13 | 山东爱特云翔信息技术有限公司（Cn） | Method, system, device and storage medium for OpenStack adaptive load balancing service |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102833355A (en) * | 2012-09-22 | 2012-12-19 | 广东电子工业研究院有限公司 | Load balance system and cloud computing oriented mechanism |
| CN103533063A (en) * | 2013-10-18 | 2014-01-22 | 北京华胜天成科技股份有限公司 | Method and device capable of realizing dynamic expansion of WEB (World Wide Web) application resource |
| CN103546542A (en) * | 2013-09-29 | 2014-01-29 | 北京航空航天大学 | Server load balancing method and device |
| CN103559072A (en) * | 2013-10-22 | 2014-02-05 | 无锡中科方德软件有限公司 | Method and system for implementing bidirectional auto scaling service of virtual machines |
-
2014
- 2014-04-22 CN CN201410163481.8A patent/CN103957246B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102833355A (en) * | 2012-09-22 | 2012-12-19 | 广东电子工业研究院有限公司 | Load balance system and cloud computing oriented mechanism |
| CN103546542A (en) * | 2013-09-29 | 2014-01-29 | 北京航空航天大学 | Server load balancing method and device |
| CN103533063A (en) * | 2013-10-18 | 2014-01-22 | 北京华胜天成科技股份有限公司 | Method and device capable of realizing dynamic expansion of WEB (World Wide Web) application resource |
| CN103559072A (en) * | 2013-10-22 | 2014-02-05 | 无锡中科方德软件有限公司 | Method and system for implementing bidirectional auto scaling service of virtual machines |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104158911A (en) * | 2014-09-01 | 2014-11-19 | 青岛海信电器股份有限公司 | Server cluster-oriented control method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103957246A (en) | 2014-07-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103957246B (en) | Dynamic load balancing method based on tenant's perception and system | |
| CN105391797B (en) | Cloud Server load-balancing method and device based on SDN | |
| US9389903B2 (en) | Method, system and apparatus for creating virtual machine | |
| CN105979009A (en) | Method for automatically balancing increased load of cloud application container | |
| WO2020133967A1 (en) | Method for scheduling shared computing resources, shared computing system, server, and storage medium | |
| CN103620578B (en) | Local cloud computing via network segmentation | |
| CN105743995A (en) | Transplantable high-available container cluster deploying and managing system and method | |
| US10171294B2 (en) | Information processing device and system design support method | |
| WO2016037479A1 (en) | Method, device and system for optimizing virtualized network function (vnf) | |
| CN104618693A (en) | Cloud computing based online processing task management method and system for monitoring video | |
| US9378039B2 (en) | Virtual machine storage replication schemes | |
| CN110661842B (en) | Resource scheduling management method, electronic equipment and storage medium | |
| CN102594861A (en) | Cloud storage system with balanced multi-server load | |
| CN103458055A (en) | Clout competing platform | |
| US11556369B2 (en) | Virtual machine deployment method and OMM virtual machine | |
| US11159604B2 (en) | Processing an operation with a plurality of processing steps | |
| EP3442201B1 (en) | Cloud platform construction method and cloud platform | |
| US10761869B2 (en) | Cloud platform construction method and cloud platform storing image files in storage backend cluster according to image file type | |
| CN103595801A (en) | Cloud computing system and real-time monitoring method for virtual machine in cloud computing system | |
| CN113630383B (en) | Edge cloud cooperation method and device | |
| CN114189525A (en) | Service request method and device and electronic equipment | |
| CN111352726B (en) | Stream data processing method and device based on containerized micro-service | |
| CN105553732A (en) | Distributed network simulation method and system | |
| US9626213B2 (en) | System and method for file injection in virtual machine configuration | |
| CN106534307B (en) | Cloud environment elastic load equalization methods based on load balancing dynamic configuration plug-in unit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |